Description

This PR defines observability requirements for components at the "Stable" stability levels. The goal is to ensure that Collector pipelines are properly observable, to help in debugging configuration issues.

Approach

• The requirements are deliberately not too specific, in order to be adaptable to each specific component, and so as to not over-burden component authors.
• After discussing it with @mx-psi, this list of requirements explicitly includes things that may end up being emitted automatically as part of the Pipeline Instrumentation RFC (RFC - Pipeline Component Telemetry #11406), with only a note at the beginning explaining that not everything may need to be implemented manually.

Feel free to share if you don't think this is the right approach for these requirements.

Link to tracking issue

Resolves #11581

Important note regarding the Pipeline Instrumentation RFC

I included this paragraph in the part about error count metrics:

The goal is to be able to easily pinpoint the source of data loss in the Collector pipeline, so this should either:

• only include errors internal to the component, or;
• allow distinguishing said errors from ones originating in an external service, or propagated from downstream Collector components.

The Pipeline Instrumentation RFC (hereafter abbreviated "PI"), once implemented, should allow monitoring component errors via the outcome attribute, which is either success or failure, depending on whether the Consumer API call returned an error.

Note that this does not work for receivers, or allow differentiating between different types of errors; for that reason, I believe additional component-specific error metrics will often still be required, but it would be nice to cover as many cases as possible automatically.

However, at the moment, errors are (usually) propagated upstream through the chain of Consume calls, so in case of error the failure state will end up applied to all components upstream of the actual source of the error. This means the PI metrics do not fit the first bullet point.

Moreover, I would argue that even post-processing the PI metrics does not reliably allow distinguishing the ultimate source of errors (the second bullet point). One simple idea is to compute consumed.items{outcome:failure} - produced.items{outcome:failure} to get the number of errors originating in a component. But this only works if output items map one-to-one to input items: if a processor or connector outputs fewer items than it consumes (because it aggregates them, or translates to a different signal type), this formula will return false positives. If these false positives are mixed with real errors from the component and/or from downstream, the situation becomes impossible to analyze by just looking at the metrics.

For these reasons, I believe we should do one of four things:

1. Change the way we use the Consumer API to no longer propagate errors, making the PI metric outcomes more precise.
   We could catch errors in whatever wrapper we already use to emit the PI metrics, log them for posterity, and simply not propagate them.
   Note that some components already more or less do this, such as the batchprocessor, but this option may in principle break components which rely on downstream errors (for retry purposes for example).
2. Keep propagating errors, but modify or extend the RFC to require distinguishing between internal and propagated errors (maybe add a third outcome value, or add another attribute).
   This could be implemented by somehow propagating additional state from one Consume call to another, allowing us to establish the first appearance of a given error value in the pipeline.
3. Loosen this requirement so that the PI metrics suffice in their current state.
4. Leave everything as-is and make component authors implement their own somewhat redundant error count metrics.